Measuring tool for code versions performance - #63
Merged
Conversation
Builds both revisions from a fresh clone into one image and runs them alternately on the same GPU, which is the only way an 8-12% difference can be told apart from the spread between two rented machines. Kept separate from the shipping image in the repository root: it has its own directory, its own build context and its own entrypoint. Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
The warning fired when both revisions measured identically with no spread, calling a quiet machine noisy. It now only fires when the repeats actually disagree, and says what that means for the comparison. Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
The benchmark picked up whatever PUBLIC_KEY held in the environment, so on a platform that sets that name for its own SSH key every run died with "public key must be 128 hexademical characters long" and the container was restarted in a loop. BENCH_PUBLIC_KEY is now the name to use, a plain PUBLIC_KEY is only honoured when it holds 128 hexadecimal characters, and the header always reports which key the run used. Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
ttl.sh was the only documented target, which forced a random image name and a lifetime of at most a day on anyone running more than one comparison. The name is now chosen to suit the registry: the local tag is reused everywhere except ttl.sh, where the tag has to carry the lifetime. Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
Co-authored-by: Gleb Alekseev <alekseev.gleb@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds
bench/, a measuring tool kept separate from the shipping image added in #62: it builds two arbitrary revisions of profanity2 into one image and runs them alternately on the same GPU.